Figure 1

Figure 1. Geographic location of Manombo Special Reserve and spatial distribution of sampled rats in a study of Angiostrongylus cantonensisrat lungworm detection in rats in southeastern Madagascar, 2022–2023. A) Location...

As part of a broader One Health study (10), we collected oral swab samples from 125 wild black rats that were captured, anesthetized, and euthanized during October 2022 and July 2023 in the Manombo Special Reserve (MSR), southeastern Madagascar (Figure 1, panels A, B). The samples were collected in 3 different habitats: littoral forest (n = 76 samples), lowland rainforest (n = 22 samples), and village-edge (n = 27 samples).

We conducted DNA extraction by using the Roche HighPure Viral Nucleic Acid Kit (Roche, https://www.roche.com). We followed our previously published pipeline for library preparation, sequencing, bioinformatic processing, quality trimming, de novo assembly, and contig screening by using standard tools and the National Center for Biotechnology Information (NCBI) mitochondrial RefSeq database (11). We annotated mitochondrial genomes by using reference genomes of A. cantonensis (NCBI Sequence Read Archive accession no. GQ398121) with manual curation.

For phylogenetic analysis, we aligned complete mitochondrial genomes with publicly available A. cantonensis sequences. We constructed a maximum-likelihood tree by using IQ-Tree 2 (12) to determine the genetic relationships of our samples and haplotype placement.

Because metagenome-assembled A. cantonensis mitogenomes were highly similar, we mapped raw reads to a representative mitogenome by using CoverM (13) with a minimum read identity of 95%. We used a read coverage threshold of >75% to determine presence or absence of A. cantonensis genomes in each sample, providing a high-confidence proxy for infection status and calculated prevalence accordingly. To identify predictors of infection, we performed a generalized linear model in R version 4.4.1 (The R Project for Statistical Computing, https://www.r-project.org), testing age, sex, location, body condition, weight, and year of sampling.

All procedures were approved by Duke University Institutional Animal Care and Use committee (approval no. A075‐23‐03), Zoo New England ethics board, and the Madagascar Ministry of the Environment (approval nos. 286/22 and 215/23/MEDD/SG/DGGE/DAPRNE/SCBE.Re). We deposited all raw sequencing data in the NCBI Sequence Read Archive. Taxonomic classification revealed that 26.26%–51.43% of reads per sample were assigned to known taxa, including bacteria (15.01%–47.72%), eukaryotes (0.53%–4.62%), and viruses (0.04%–18.78%). Viral sequences mainly comprised anelloviruses and bacteriophages (10).

Figure 2

Figure 2. Mitochondrial genome organization and sequence coverage maps of Angiostrongylus cantonensis rat lungworm from black rat oral swab specimens in southeastern Madagascar, 2022–2023.

Figure 3

Figure 3. Whole-mitogenome maximum likelihood phylogenetic tree of complete sequences available for Angiostrongylus cantonensis rat lungworm and 3 related Angiostrongylus species along with the sequences identified in study of ...

We detected 5 complete A. cantonensis mitochondrial genomes from black rat oral swab specimens, de novo assembled with high coverage (GenBank accession nos. PX571103–7) (Figure 2). All 5 mitochondrial genomes were 13,503 nt in length with 26.8% guanine and cytosine content, in comparison with published A. cantonensis mitogenomes that were 13,497–13,519 nt in length with 26.7%–26.8% guanine and cytosine content. The mitogenomes from Madagascar revealed 99.96%–100% sequence similarity among themselves and 96.44%–99.99% similarity compared with global sequences. Of note, the genomes from Madagascar demonstrated >99.98% similarity to sequences from Valencia, Spain (GenBank accession no. PP748572). Phylogenetic analysis placed all sequences from Madagascar within clade II, grouping with the Val-II haplotype and clustering with the haplotype Ac8 from Brazil (Figure 3) (14).

A. cantonensis prevalence in black rat oral swabs was determined to be 10.4% (95% CI 6.2%–16.9%; 13/125 samples). All positive rats were from littoral forest habitats from the littoral forest of the MSR (Figure 1, panel B). Positive cases were detected in both sampling years, with no significant difference in prevalence between years: 9.67% (95% CI 4.5%–19.6%; 6/62) in 2022 and 11.11% (95% CI 5.5%–21.4%; 7/63) in 2023. Logistic regression analysis identified age as the only significant predictor of infection. Young rats had significantly lower odds of infection compared to adults (odds ratio 0.13, 95% CI 0.02–0.77; p = 0.03) (Table). We observed no significant associations for sex, body condition, weight, or year of sampling (Table).

The 5 mitogenomes we obtained were nearly identical to each other (99.96%–100% similarity), indicating a highly conserved parasite population within the MSR littoral forest and low genetic diversity of A. cantonensis lungworm in Madagascar, which is similar to another report (3). Phylogenetic analysis further demonstrated that all the sequences cluster within clade II and specifically group with the Val-II haplotype, previously reported in Europe and associated with haplogroup Ac8 from Brazil (14). That pattern supports the hypothesis of recent or historical introductions mediated by human activities, particularly through the movement of invasive rats and intermediate hosts via trade and transport networks (3,14).

The overall prevalence in rats in the current study was 10.4% (95% CI 6.2%–16.9%; 13/125 rats), higher than the 2.7% prevalence previously reported in rats from central-eastern Madagascar (n = 78) (9). All positive rats were from the littoral forests. Host-related factors also contributed to infection patterns. Age was the only significant predictor; young rats showed markedly lower odds of infection compared with adults. That finding reflects cumulative exposure over time, because older rats have more opportunities to encounter infected intermediate or paratenic hosts (15). We observed no significant effects for sex, body condition, weight, or year, suggesting that exposure-driven processes might be more necessary than intrinsic host factors in shaping infection risk. No positive detections were observed in the lowland rainforest or village-edge habitats, potentially because of intermediate host distribution (5).

This study provides metagenomic-based mitochondrial genome characterization and prevalence estimates of A. cantonensis lungworm from black rats in southeastern Madagascar. Because of the parasite’s life cycle, oral swab specimens likely capture first-stage larvae expelled from the respiratory tract, which are known to be present in the oral cavity during patent infections. However, because necropsy data were not systematically collected for parasitologic examination, we could not assess the relationship between larval detection in oral swab specimens and adult worm burden in the pulmonary vasculature. Future studies combining oral swab specimens with necropsy or quantitative parasitologic methods would help validate this relationship. Nevertheless, our findings demonstrate that complete A. cantonensis mitochondrial genomes can be recovered from oral swab specimens, supporting their use as a complementary, minimally invasive approach for molecular detection and genomic characterization of this parasite.